AMD confirms that the HP Envy x360 uses Mobile XFR

AMD's Ryzen Mobile APUs are some of the most exciting products to launch in the PC space this year, and as enthusiasts, we wanted to channel that excitement by getting a review of one of those systems for our readers as quickly as possible. Our initial results showed promising performance for the chip, but some potentially less-promising numbers for battery life.

We've been trying to get to the bottom of those figures as we spend more time with the system, and in the course of that work, we have learned that one fundamental premise of our initial review was incorrect. AMD got in touch this morning regarding our results and officially confirmed that the Ryzen 5 2500U inside the Envy x360 we tested uses Mobile XFR, which is to say that it has a 25W TDP instead of the 15W envelope that was widely reported ahead of the system's arrival on the market. Here's the official line from the reviewer's guide:

Not all notebooks with the AMD Ryzen™ mobile APU will offer the necessary thermal solution to enable the performance upside of mXFR, but the HP ENVY x360 featuring the AMD Ryzen™ Processor with Radeon™ Vega Graphics is the first solution to do so. Users will look for “amplified mXFR performance” in the marketplace should they desire a laptop that offers this capability.

That new information significantly affects our analysis of the performance and performance-per-watt characteristics of the Envy x360 versus the 15W implementations of the Core i5-8250U that we tested it against, and I'll be incorporating this information into some updates throughout my original article as quickly as I'm able.

From AMD's reviewer's guide and in conversations with David Kanter for an upcoming podcast, though, we know that Mobile XFR primarily affects sustained workloads, and it probably means that the Ryzen 5 2500U may only achieve performance parity or fall somewhat behind the i5-8250U in multithreaded workloads in a 15W implementation. We really won't know until systems configured to use 15W versions of this APU show up for us to test. For now, buyers will want to take this new information into account as they weigh the Ryzen-powered version of the Envy x360 against the competition.

Comments closed
    • odizzido
    • 2 years ago

    Thanks for correcting this.

    • IGTrading
    • 2 years ago

    NoteBookCheck said the CPU uses less energy thab Kaby Lake based competition in 90% of the cases.

    In my opinion, the CPU itself should not be blamed because the iGPU is in there as well.

    So on Intel and nVIDIA laptops you have 15W for the CPU and 25W for the GPU with a total of 40W.

    Therefore AMD’s default 15W TDP with the mXFR reving up to a total 25W TDP is VERY impressive.

    It is my opinion that GF MX150 would have have such high results if it would have to fit into a tiny 10W TDP.

    Besides, this is the absolute first time we see the Raven Ridge in action with the absolute first Vega 8 driver.

    The capabilities and efficiency will improve very soon with newer drivers and OS updates.

      • Jeff Kampman
      • 2 years ago

      Low energy consumption != higher efficiency.

    • Bumper
    • 2 years ago

    so apparently it [i<]is[/i<] the HP model that's the problem. using the same model the amd and intel version are only separated by 38 minutes in their battery test. [url<]https://www.laptopmag.com/reviews/laptops/hp-envy-x360-15t[/url<] on a side note... someone disparaged me about Lenovo efficiency in a previous thread.. damn look how well my yoga 720 does in this review. best battery life...

      • Mr Bill
      • 2 years ago

      Don’t neg him ya mean things. He is correct… [quote<]HP's Envy notebooks offer similar performance, with speed for solid multitasking. The AMD version of the Envy x360 we tested features a Ryzen 5 2500U CPU and 8GB of RAM, while the Intel-based Envy rocks an Intel Core i7-8550U with 16GB of RAM.[/quote<][quote<]The Laptop Mag Battery Test (web surfing at 100 nits) drained the Intel model of its charge in 5 hours and 49 minutes, while the AMD model hit empty after 5 hours and 11 minutes.[/quote<]

        • DancinJack
        • 2 years ago

        As if being correct always stopped down votes in these comments. Funny.

          • Growler
          • 2 years ago

          Your “facts” mean nothing if I believe something hard enough!

          That’s why I’m not a 41 year old single guy. I’m young, handsome, and fighting off the attention of many beautiful women constantly. Reality just gets in the way.

      • DavidC1
      • 2 years ago

      That Yoga 720 is not stellar considering the battery capacity. It’s at 72WHr. That Yoga 720 tested at Laptopmag is a 15-inch screen version.

      • tsk
      • 2 years ago

      Interesting, I think testing the same model laptop with the different CPUs is the best way to compare Intel and AMD, there’s just too many variables otherwise.

    • Unknown-Error
    • 2 years ago

    So it is 25w AMD vs 15w Intel? Doh!!!

    So much competitivenes. At 15w the CPU benchmarks of the 8250U is insane. Even the 7200U suddenly looks good.

    Mediocre CPU, horrible battery life? AMD will have better sales but only compared to bulldozer.

      • neblogai
      • 2 years ago

      At ’15W’, 8250U laptops can easily use 40W. 7200U usually settle closer to 30-35W. It all depends on cooling and setup.

        • Klimax
        • 2 years ago

        I don’t think you improved comparison much…

        • DavidC1
        • 2 years ago

        Yea but that doesn’t mean it uses it sustained. It’s not on all the time you know.

          • Kraaketaer
          • 2 years ago

          Exactly. The recent NotebookCheck review of the Surface Book 2 puts this nicely into perspective – the i7 8550U peaks at ~35W (even with passive cooling!), then “throttles” to ~25W, then goes into a thermal slowdown loop which makes it (over time) average at ~16W under continous load. IIRC it drops somewhat below base clocks (~.1GHz) during the low points of the loop, but that’s negligible. It seems that OEMs have some more control over what limits the CPU than they used to – whether it’s thermals, power, or both.

          The key point here is sustained performance. CPU manufacturers realizing that even a weak cooling setup can handle ~30-40W for short bursts is a huge win for mobile performance and useability, which we should all be happy about.

          From my perspective, it looks like Ryzen Mobile is quite competitive, though perhaps not entirely up there with Intel. Then again, the effect of the (vastly superior) iGPU is hard to measure. Ryzen desktop is more efficient than KBL, but the 14++ node does seem to have made some truly meaningful strides in efficiency. What impresses me the most is that we now get 15W quad core chips with the same base clock as 17W dual core chips in the Haswell era, only turbo clocks are more than 1GHz better. This is quite amazing, IMO. And AMD suddenly being on roughly the same level? That’s nigh on miraculous.

      • Mr Bill
      • 2 years ago

      If you do the math, both machines were drawing less than 8W for the test. So, TDP does not matter.

    • Kretschmer
    • 2 years ago

    Ahh, there’s the mobile AMD parts that we know.

    • maroon1
    • 2 years ago

    LOL 25w TDP and still loses to i5 8250U in most CPU tests

    I wonder how well it will perform in 15w TDP mode

      • Klimax
      • 2 years ago

      Overclocked Apollo?

      • shank15217
      • 2 years ago

      Ah yes, because tdp = power usage …

    • tsk
    • 2 years ago

    I’m a bit confused so mobile XFR =cTDP up?

      • chuckula
      • 2 years ago

      More like LEVEL UP!

    • mtruchado
    • 2 years ago

    Jeff, please, let me know how easy/difficult is to update that system, this is, to upgrate It to have an SSD and if the memory is soldered

    Thanks.

    Edit:

    HotHardware published already pictures of this machine opened (credits to Jeff for the link)

    [url<]https://hothardware.com/reviews/amd-ryzen-5-2500u-benchmarks-revisited-hp-envy-x360-ssd-update[/url<] As you can see, 1x SATA and 1x nvme m2 slots are there, plus 2x DDR4 ram sockets.

      • DPete27
      • 2 years ago

      [quote=”from the review article”<]To make benchmarking the Envy bearable, I stuck a Samsung 960 EVO 500GB NVMe from our stockpile inside.[/quote<]

        • mtruchado
        • 2 years ago

        I read that, but I wanted to know how difficult is to achieve 🙂

      • Jeff Kampman
      • 2 years ago

      It’s tough to get inside (you need to partially or fully remove one of the rubber feet on the underside of the machine, and HP lists it as a consumable), but once you’re in there, everything is socketed. I only had to borrow an M.2 hold-down screw from my stockpile to secure the SSD.

        • DPete27
        • 2 years ago

        [quote<]one of the rubber feet on the underside of the machine, and HP lists it as a consumable[/quote<] Edible computer parts?! I think we just discovered the successor fad to RGB LEDs

    • chuckula
    • 2 years ago

    Do I think that Intel’s not-particularly-high end IGP from 2015 that’s in Kaby Lake R would beat RyZen if the chip is given a 66% larger power envelope? No.

    Do I think that maybe some of the hype over Ryzen’s IGP is starting to wear off as the real-world tests start to come out in greater detail? Absolutely.

      • neblogai
      • 2 years ago

      That is for the best. Raven Ridge is a great CPU, which also comes with free iGPU. If considered for gaming- it should be compared to competition directly:
      Intel iGPU vs AMD iGPU or
      Intel + dGPU v AMD + dGPU.
      In both instances, RR should be cheaper, and offer better price/performance.
      iGPU should only be considered as a fallback, or budget feature: important for bare laptops with cheap RR SKUs with cut-down CPU cores- it can be extremely competitive there. And in desktops- it can push out both i3s and GT1030- which are also pure budget segment. But using full RR just as iGPU would be a waste of a great quadcore CPU that at 15-25W is able to drive way more powerful graphic cards.

    • DragonDaddyBear
    • 2 years ago

    Any way to configure the TDP in the BIOS/UEFI?

      • Jeff Kampman
      • 2 years ago

      Not yet.

    • DPete27
    • 2 years ago

    Just think, if AMD could improve their voltage algorithm they could get the same performance as XFR but within that 15W limit.

      • smilingcrow
      • 2 years ago

      Are you being sarcastic? I hope so.

    • ET3D
    • 2 years ago

    Another point raised by NotebookCheck was that performance on mains power is significantly higher than on battery. This is normal, but for me what matters is gaming on battery.

    So are the benchmarks done on battery or mains, and if mains, is it possible to also include battery powered results?

      • DragonDaddyBear
      • 2 years ago

      Did they mention what the power profile was? It can change based on wall or battery power.

    • xeridea
    • 2 years ago

    Can XFR be disabled in UEFI, or software? I know they are locked chips, being mobile and all, but it would be a power saving feature, so should be allowed, sacrificing a small amount of performance for big battery life gain.

      • Jeff Kampman
      • 2 years ago

      Not yet.

        • DPete27
        • 2 years ago

        I don’t have any experience with AMD laptops, but my HTPC has an AMD APU. Could you use Wattman to circumvent?

        Come to think of it, I’ve used Windows power settings to disable Intel turbo boost. Simply set “max processor state” to less than 100%.

      • MOSFET
      • 2 years ago

      Just don’t do anything that requires turbo.

      • Mr Bill
      • 2 years ago

      Its pretty clear from derFunkenstein’s post of the NotebookCheck review that both the TR tested systems were running well under their respective TDP for the web browsing battery life benchmark.
      The Ryzen 5 2500U Vega8 combo is pulling 7.69W.
      The Acer Swift 3 i5-8250U MX150 combo is pulling only 3.14W.

      • DavidC1
      • 2 years ago

      xeridea:

      We’ve come a long way from the days where the CPUs run at fixed frequency, power and voltage states.

      Rather, they quickly go from one state to another depending on what you do. Turbo modes can increase power if you are doing something that requires a constant load, but you aren’t doing that when web browsing. Clocks burst up for a short period of time when a page loads, and goes down in ms times after its done.

      For a while a concept called Hurry Up and Get Idle, or HUGI was used to develop notebook platforms. The idea is a faster execution allows the CPU to go to idle states faster to save power.

      So NO, disabling XFR won’t impact battery life in web browsing and video playback scenarios.

    • derFunkenstein
    • 2 years ago

    That’s consistent with NotebookCheck’s review, which also mentioned a 25W TDP in their conclusion:

    [url<]https://www.notebookcheck.net/HP-Envy-x360-15-Ryzen-5-2500U-Radeon-Vega-8-Laptop-Review.266614.0.html[/url<] [quote<]Since this is the first commercially available notebook equipped with Raven Ridge, our Verdict can be split into two parts. The first part relates to the AMD hardware itself. On paper, the [b<]25 W[/b<] cTDP Ryzen 5 2500U APU and RX Vega 8 GPU are able to stand neck-to-neck with current 15 W Kaby Lake-R options (i5-8250, i7-8550U) while outperforming even the Iris Pro Graphics 580 in synthetic benchmarks. [/quote<] (emphasis mine)

      • derFunkenstein
      • 2 years ago

      BTW, Jeff, that review mentions tons of random crashes and slowness. You apparently didn’t experience that, right?

        • Jeff Kampman
        • 2 years ago

        Nope, we only used a fresh Windows install with the bare minimum of drivers needed to keep the machine functioning properly. No stability issues.

          • derFunkenstein
          • 2 years ago

          Ah, that’s probably the difference. I’m suppose NotebookCheck probably used the stock Windows installation.

          • dragosmp
          • 2 years ago

          …and an SSD. I’m not saying an HDD is unstable, just that a bloated win install and a HDD may lead to slowdowns.

            • derFunkenstein
            • 2 years ago

            I had also assumed the slowdowns were due to the hard drive, but it could also be HP’s bloatware.

            • dragosmp
            • 2 years ago

            I get their POW, as this is how most people would run their PC. Us on the other hand would do exactly what Jeff did, which makes this article very useful. If the only review of this HP lappy would be the NC review, I wouldn’t touch it.

            I’m curious about the crashes they had – white screen freezes and games not starting – it seems like driver issues. Would a stock install be as bad? I owned may Radeon GPUs and they’ve always been dodgy with some drivers, but good with others. Jeff didn’t mention crashes, maybe NC either tested more apps, so found more bugs, or the stock install is a disaster.

            • dragontamer5788
            • 2 years ago

            There’s a known Radeon / Overwatch bug at the moment. But at the same time, HP’s “Touchpoint” software has been causing slowdowns and issues.

            It seems like Jeff was able to avoid the issues of “HP Touchpoint” by reformatting the machine.

            • derFunkenstein
            • 2 years ago

            Oh, that’s for sure. Almost nobody is reformatting these things when they get them. Just another incompetent OEM bundling vaguely useless software ruining the show here.

      • Mr Bill
      • 2 years ago

      Interesting that their “load” test at 77 minutes reflects pretty closely the 45W draw for Witcher while the 398 minutes for “WiFi v1.3” reflects only an 8.2W draw. The TR’s result is drawing 54.5W-hr / 7.08 hr = 7.7W usage without the screen. Video browsing really does not take much power.

      • Mr Bill
      • 2 years ago

      Furthermore, the Acer Swift 3 (i5-8250U, MX150) is pulling 45.7W-hr / 14.55 hrs = 3.14W on the web browsing benchmark. So, the 8250U IGP is very efficient compared to the Ryzen/Vega8 combo. But if in gaming, the wattage of an (i5-8250U, MX150) system like the one in your link e.g. [quote<]The Xiaomi Mi Pro is perhaps the best example of this as it carries the i5-8350U CPU, MX150 GPU, and the same size 1080p IPS display as our Ryzen HP. When subjected to average (3DMark06) and Witcher 3 loads, both systems are nearly identical at 46 W to 51 W each. [/quote<] The (i5-8250U, MX150) may run faster than the Ryzen/Vega8 combo but power consumption is about the same when gaming.

    • NTMBK
    • 2 years ago

    Ah damn, that makes the gaming benchmarks look a lot less impressive. I wonder how different they would look without XFR.

      • cynan
      • 2 years ago

      I still think the performance/TDP is fairly impressive. It almost competes with the MX150, and that’s a Pascal chip (efficient) which has a TDP of 25W alone (ie, not counting the CPU).

        • maroon1
        • 2 years ago

        [url<]https://techreport.com/r.x/2017_11_25_AMD_s_Ryzen_5_2500U_APU_reviewed/Dota2BL_average_fps.png[/url<] [url<]https://techreport.com/r.x/2017_11_25_AMD_s_Ryzen_5_2500U_APU_reviewed/RL1080_average_fps.png[/url<] [url<]https://techreport.com/r.x/2017_11_25_AMD_s_Ryzen_5_2500U_APU_reviewed/W3900_average_fps.png[/url<] [url<]https://techreport.com/r.x/2017_11_25_AMD_s_Ryzen_5_2500U_APU_reviewed/Doom_average_fps.png[/url<] MX150 is 70% faster in Dota and 57.5% faster in Doom (vulkan). It does not look like "almost competes" .The gap is smaller in other two games I linked, but it still over 25% Also, [url<]https://hothardware.com/ContentImages/Article/2690/content/Far-Cry-2-Ryzen-Mobile-Benchmark-Result2.png[/url<] [url<]https://hothardware.com/ContentImages/Article/2690/content/GRID-Autosport-Ryzen-Mobile-Results4.png[/url<] The gap is even bigger in older games. 71% faster in far Cry 2 and 81% faster in GRID autosport (it is also worth to mention that MX150 is more than double the minimum fps in GRID)

          • cynan
          • 2 years ago

          Of course mobile Ryzen “almost” (but not actually) competes when the only other option for an IGP is the significantly slower 620 on Intel’s chips.

          And it competes quite closely with the old version (940mx) that was the standard option for thin laptops sold up until a month or two ago that had the same TDP.

          The impressive part is that AMD is managing to compete on a performance per TDP basis with both Nvidia and Intel.

          Don’t be maroon!

        • DavidC1
        • 2 years ago

        “It almost competes with the MX150, and that’s a Pascal chip (efficient) which has a TDP of 25W alone (ie, not counting the CPU).”

        I do agree the 3D performance level is impressive, considering its on the Iris Pro level.

        But, what you are saying isn’t necessarily true. On systems with relatively low end GPUs, the CPU isn’t being utilized fully. So it may not be at 15W.

        When you see the results here, you can see the load power consumption numbers are similar to the MX150 + KBL-R configuration: [url<]https://www.notebookcheck.net/HP-Envy-x360-15-Ryzen-5-2500U-Radeon-Vega-8-Laptop-Review.266614.0.html#toc-energy-management[/url<] The MX150+8250U Xiaomi shows 3.3W higher average load power than the Ryzen 2500U system. Of course certain MX150 systems show higher average load power.

    • chuckula
    • 2 years ago

    [quote<]AMD got in touch this morning regarding our results and officially confirmed that the Ryzen 5 2500U inside the Envy x360 we tested uses Mobile XFR, which is to say that it has a 25W TDP instead of the 15W envelope that was widely reported ahead of the system's arrival on the market.[/quote<] That explains some of what we have seen but not all of it considering even a higher-TDP part will still clock down quite heavily in lighter workloads like TR's web bench or a video playback that's using the acceleration hardware. Bonus points for the anti-TR conspiracy theorists who accused TR of lying about the power consumption levels of Intel's mobile parts being literally 180-degrees wrong too.

Pin It on Pinterest

Share This